The present invention relates generally to the field of voice recognition, and more particularly to voice pattern coding and catalog matching.
Many computing devices include voice, or speech, interfaces to assist users in interfacing with the computing devices. Currently, voice recognition software applications can use model speech patterns/formulas which can be attained by various means. In some cases, interfaces are command driven, where users can speak a predefined set of commands, and in other cases, interfaces allow free form speech. Manufacturers or providers of voice interfaces typically pre-set one language as a default, but provide the user with the option to change the primary speaking language if needed.
Known speech interfaces may include automatic speech recognition (ASR) engines which support multiple language recognition capabilities with a predefined set of assumptions. Typical ASR engines require training time for learning a user's vocabulary, where the engines are trained to recognize words spoken by the user.